JPH0519354B2 - - Google Patents

Description

[Detailed description of the invention]

INDUSTRIAL APPLICATION FIELD The present invention relates to an image signal processing method for performing gradation correction of an image signal obtained by reading a gradation image such as a photograph by photoelectric scanning. Conventional configuration and its problems In a device that processes a read image signal obtained by photoelectrically scanning a gradation original such as a photograph, such as an image scanner device or an imaging device, the read image signal is transferred to an image reading element or an image sensor. Various image signal processing circuits are provided for the purpose of matching the characteristics of recording or display elements and for the purpose of artificially processing the quality of images. In particular, many of recent image processing apparatuses have a function of artificially converting gradation characteristics. FIG. 1 shows processing of read image signal processing in a conventional image signal processing device.
In particular, the flow of processing related to gradation conversion is shown. In the conventional example shown in FIG. 1, the read image signal v ₁ , which is inversely proportional to the index of the image density of the original, is an image signal amplified by the preamplifier 1 so that the amplitude varies within a predetermined range. Generate v ₂ . The image signal v ₂ is converted into an image signal by a range amplifier 2 that artificially moves the gradation in a straight line.
Convert it to _v3 . The image signal v ₃ is converted into an image signal v ₄ by a gradation converter 3 that artificially converts the gradation. The console 4 has a function of generating control signals for performing artificial gradation conversion, and provides a white level signal _c1 and a black level signal _c2 to the range amplifier 2, as shown in FIG. Image signal conversion is performed using the input/output characteristics as follows. The gradation converter 3 has a function of converting the input image signal v ₃ using an appropriate function F and converting the output image signal v ₄ so that v ₄ =F(v ₃ ). This is called a key correction curve, and an example thereof is shown in FIG. Control signals for highlight level c ₃ , middle tone level c ₄ and shadow level c ₅ are applied to the gradation converter 3 from the console 4 to control the highlight area 5, middle tone area 6 and shadow in FIG. The slopes of straight lines 8, 9, and 10 in each region of region 7 are controlled. As a method of controlling the slope of such a straight line, a method of varying the load resistance or feedback resistance of an amplifier is often used. The gradation correction curve in FIG. 3 shows that the function F is given by a broken line function consisting of straight lines 8, 9, and 10, but the boundaries of the normal areas 5, 6, and 7 are predetermined. The classification is based on semi-fixed values. In the conventional example described above, the gradation converter 3 simply divides the level of the input image signal _v3 into three areas of highlight, middle tone, and shadow and performs signal conversion processing. It is only possible to give a gradation correction curve that is accurate. Furthermore, since the density of the original image is determined by human visual judgment when specifying the artificial gradation correction curve, the coordinate axes in Figure 3 that determine the function F of the gradation correction curve are based on density linearity. Coordinates are more ergonomic. However, in the case of an image signal proportional to the reflectance of the original image as shown in Figure 1,
The coordinate axes in Figure 3 are reflection straightness coordinates, and it is difficult to set the function F of the tone correction curve by intuitive judgment of the original image, and normally determining such a tone correction curve is a task that requires considerable skill. There is. Purpose of the Invention The present invention has been made in order to solve the above-mentioned problems in tone conversion processing for image signals proportional to reflectance. The present invention aims to provide a signal processing method that can easily process even complex gradation transformations. Structure of the Invention In order to achieve this object, the present invention obtains a digital input image signal by digitally converting a signal obtained by amplifying the amplitude range of an image signal proportional to the amount of reflected light of an original image to a predetermined level. , create data indicating a gradation conversion curve based on the coordinate values of multiple points on the gradation correction curve drawn on the normalized input density vs. normalized output density coordinate plane and the values of white density and black density, and digitally For an image signal processing device that converts the gradation of an input image signal using data representing a gradation conversion curve to obtain a digital output image signal, the procedure for creating the data of the gradation conversion curve involves converting the digital input image signal into a digital output image signal. N
The maximum digital output image signal level is M, the input white density and black density values are D _W and D _B , respectively, and the coordinate values of multiple points P _o on the gradation correction curve are (x _o , y _o ) (N, n are natural numbers), the digital input image signal x _iI vs. digital output image signal v _pI The coordinate value of point P _o ′ on the coordinate plane (x _o ′, y _o ′)
is obtained from the relational expression x ₁ ′=2 ^N −1/1−10 ^DW-DB {1−10 ( ^DW-DB )×xI} y ₁ ′=M While sequentially updating the level v _iI of the input image signal from 0 to 2-1,
Compare the coordinate value x _o ′ of multiple points P _o ′ to find J such that x _J ′v _iI x _J+1 ′ (J=0 to n), and
From the function F _J that indicates a straight line between P _J ′ and point P _J ′ ₊₁ , the tone-converted digital output image for the digital input image signal v _iI is calculated from the functional formula v _pI = F _I (v _iI ). This is to obtain the signal v _pI . DESCRIPTION OF EMBODIMENTS First, an apparatus for realizing the present invention will be described with reference to the drawings. FIG. 4 is a schematic diagram of the device configuration showing one embodiment of the image signal processing device. In this embodiment, the output picture signal _v3 of the range amplifier 2 is converted by the A/D converter 11 into an 8-bit parallel digital picture signal _v5 .
In the mode that converts the gradation of the image signal,
The digital image signal v ₅ is directly connected to the address signal v ₆ of the gradation conversion memory 13 by the address selector 12, and is a 256-byte area in which image signal conversion values obtained from a preset gradation correction curve are stored. An 8-bit parallel digital image signal _v7 whose gradation is converted by a so-called memory reference method is output, which is used as an address signal for reading data from the gradation conversion memory 13 consisting of the following. By configuring a tone converter using a digital image signal and a reference memory in this way, it is possible to easily perform data conversion according to a complicated correction curve as described later. Now, as pointed out in the conventional example, it is ergonomic to set the gradation correction curve in a coordinate system proportional to density even for an image signal processing path proportional to reflectance.
In this case, in order to make it possible to set the function F of the gradation correction curve with a value proportional to the density, a microprocessor is used to perform the coordinate transformation calculation. The output image signal _v3 of the range amplifier 2 calculates the density of the original image when the white density (D _W ) and black density (D _B ) are given.
If D _X It is expressed as The tone correction curve for such an input image signal is set by setting the white density (D _W ) to 0 and the black density (D _B ) to 1, as shown in FIG. The normalized input density coordinate axis (x axis) and the normalized output density coordinate axis (y
It is best to give it on the plane of the axis). x on the normalized input coordinate axis has the following relationship with the original image density _DX : x = ( _DX − D _W )/( _DB − D _W ),
Therefore, x _c on the x-axis in FIG. 5, which corresponds to the image signal level of the input image signal v ₃ , can be calculated using equation (1) as x _c = 1/D _B-DW log [1-v ₃ { 1-10 ^-(DB-DW )}〕……(
2) Determined by more. As shown in FIG. 5, this gradation correction curve is given by a broken line function of a plurality of points P ₁ to _{P o} (up to P _o =4 in the figure). On the other hand, the 8-bit parallel digital image signal corresponding to the image signal level of 0 to 255 of the 8-bit parallel digital image signal _v5 is sent to the gradation conversion memory 13 in FIG.
The microprocessor 14 is used to obtain the signal level value of _v7 from the gradation correction curve. The x and y coordinate values of multiple points P ₁ to _{P o} on the gradation correction curve in Fig. 5, (x ₁ , y ₁ ), (y ₂ , y ₂ ), ... (x _o , y _o ) are The data is inputted from the console 15 to the microprocessor 14 via the interface 16 and stored in the storage area 17 within the microprocessor 14 in sequence. In addition, the values of white density (D _W ) and black density (D _B ) sent to the range amplifier 2 are also stored from the console 15 in the storage area 18 in the microprocessor 14, and c ₁ = 10 ^{- DW} c ₂ = 10. ^-DB (3) After performing data conversion, the white level signal c ₁ and black level signal c ₂ are transmitted to the range amplifier 2 via the interface 19. Since the above gradation correction curve is given in the normalized density coordinate system, the digital image signal v ₅ is proportional to the reflectance.
The image signal level value of the digital image signal _V7 after conversion processing corresponding to each level of 0 to 255 is calculated by converting the gradation correction curve in Figure 5 to the output image signal level versus input image signal level as shown in Figure 6. It must be determined from the gradation transformation curve transformed onto the coordinate axis plane. Points in Figure 6
P ₁ ′ to _{P 4} ′ correspond to P ₁ to _{P 4} in FIG. 5, and the procedure for obtaining such a correction curve is to calculate v ₃ from any coordinate point v _i on the = v _i /255, obtain the normalized input density value x _c from equation (2), and use Fig. 5 to calculate the normalized output density y _c as a gradation correction curve expressed by a broken line function. Then, by calculating v _p =255×y _c to an integer, the obtained value is placed on the coordinate axis plane of FIG. 6, and a gradation conversion curve is obtained. In order to perform the above calculations, the microprocessor 14 executes the operations according to the processing flowchart shown in FIG.
It is temporarily stored in the storage area 20 inside. Figure 8 shows the storage areas 17, 18, 20 used in the flowchart of the process in Figure 7.
FIG. 8a shows an example of the arrangement of each area in the storage area 17.
array M ₁ (), b is array M ₂ of storage area 18
() and c exemplifies the array M ₃ () of the storage area 20. Array M ₁ () is the input coordinate point in Figure 5.
Examples of storage locations of coordinate values of P ₁ to P _o and storage locations of slope u o and y-axis intercept W _I of the equation y=u _o x + W _o representing the straight line between points P _I and P _{I+ 1} , n indicates the number of input coordinate points, and the procedure for determining the slope u _o and the intercept W _I is shown in step 1 in the processing flow chart of FIG.
In step 2 following step 1 in FIG. 7, the signal level v ₁ of 0 to 255 of the input image signal v ₅ to the gradation conversion memory 13 in FIG. 4 and the image signal level v ₃ in equation (2) are Since the relationship is v ₃ = v _i /255, the displacement constant ΔK = 1/
After setting the input image signal level _v3 using 255, find the normalized input density of the corresponding point _xc on the x axis in Figure 5, and calculate the gradation correction curve in any straight line area in the figure. Search for the approximation, select the slope and intercept values of the linear approximation equation calculated in step 1, obtain the normalized output concentration y _c , and after converting it into an integer in the range of 0 to 255, the result is shown in Figure 4. In the storage area 20 of FIG. 8c, the array M ₃ ()
It is stored as gradation conversion data like this. The gradation conversion data obtained in steps 1 and 2 as described above are sequentially written into the gradation conversion memory 13 via the interface 21 from the microprocessor 14 shown in FIG. The control signal c ₆ in Fig. 4 (hereinafter,
The timing chart of control signals, output signals, etc. is shown in FIG. 9 and will be explained) is a read/write mode switching signal for the gradation conversion memory 13. In the write mode, the address selector 12 is connected to the counting pulse c ₇ from the interface 21. The output signal _c8 of the address counter 22, which counts 0 to 255 and generates an address signal _c8 with a value of 0 to 255, is connected to the address line of the gradation conversion memory 13, and further connected to the sampling timing generator of the A/D converter 11. 23 is held in a dormant state.
The address counter 22 sets the count value to 0 at the first count pulse _c7 after the mode switching signal _c6 changes to the write mode, and the address counter 22 sets the count value to 0 in the memory area 2 of the microprocessor 14.
The gradation conversion data v _pI is sent one data at a time to the interface 2 according to the order of the array M ₃ () from 0 to c in Figure 8.
1 to the data line c ₉ , followed by the signal line
Step 3 shown in FIG. 7, in which the write clock to the gradation conversion memory 13 is sent out at step _c10 , is repeated. When the processing flow shown in FIG. 7 is processed by a microprocessor having the calculation speed shown in Table 1, Step 1 requires 0.05 seconds and Step 2 requires 5.05 seconds. Since step 3 is simply a 256-byte data transfer, the impact on processing time is extremely small. As is clear from Table 1, the operations that require the most time in Steps 1 and 2 are exponent and logarithm operations, with Step 2 in particular requiring 256 logarithm operations. However, in the case of recent high-speed image scanning devices, etc., the time required for the calculation processing of the gradation correction curve as described above cannot be ignored, and even faster processing is desired.

[Table] In order to achieve these objectives, the present invention:
In obtaining the gradation conversion curve of FIG. 6 from the inputted gradation correction curve of FIG. 5, the processing method shown in FIG. 10, which is different from the processing method of FIG. 7, was used. The principle of this processing method is that only the coordinates of multiple points P ₁ to _{P o} that define the gradation correction curve in Figure 5 are placed on the output image signal level vs. input image signal level coordinate plane in Figure 6.
The purpose is to use a broken line function consisting of a group of straight lines arranged as points P ₁ ′ to _{P o} ′ and a straight line approximation between points P _I ′ and P _I+1 ′ as a gradation conversion curve. This is indicated by the dashed dotted line. In this processing method, the sixth
Compared to the processing method shown in FIG. 7, some errors occur, such as points P ₁ ' to _{P 2} ' in the figure. However, the image density of reflective originals that are usually handled has a black density (D _B ) of about 2.0, and the density difference indicated by the x coordinates of adjacent input points (x _{I +1} − x _I ) + (D _B −
D _W )<2.0/5 or less, the above error has almost no effect on tone conversion. In the processing method shown in FIG. 10, the array related to the storage area 17 installed in the microprocessor 14 shown in FIG. 4 is set as the array M ₁ () shown in FIG.
In, x′=255/1−10 _(DB-DW) {1−10 ^-(DB-DW).xI }……(4
) Find x _o ′ from the relational expression and store it in a predetermined position of the array M ₁ (I), u _I ′=255(y _I+1 −y _I )/x _I+1 ′−x _I ′ ……( Find u _o ′ from the relational expression in 5) and store it in a predetermined position in the array M ₁ ′(I), then w _I ′=255y _I −u _I ′×x _I ′ ...... From the relational expression in (6), w _o ' is determined and stored at a predetermined position in the array M ₁ (I). In step 2' in FIG. 10, the input image signal level v _i in FIG _. ′ and the intercept w _J ′ are determined, and the output image signal level v _pI is determined using the relational expression v _pI = u _J ′ + u _i + w _{J ′} ...(7). Store _pI sequentially. After calculating the 257 output image signal levels, the 256-byte gradation conversion data is gradation-converted from the storage area 20 in the microprocessor 14 in FIG. 4 in exactly the same manner as step 3 in FIG. Transfer to memory 13. By using such a second processing method, when the number of input points n of the gradation correction curve is 20, the time can be shortened to 1.2 seconds. Furthermore, in the case of having the configuration shown in FIG. 4, by making the storage area 17 in the microprocessor 14 a non-volatile storage area configuration such as a magnetic recording device,
_A plurality of gradation correction curves as shown in FIG _.
4, it can be stored for a long time, and a code for curve matching, such as a number for each gradation correction curve, can be input from the console 15, and a set of already saved coordinate values of P ₁ to _{P o} can be saved. It is also possible to easily equip the microprocessor 14 in advance with a procedure for generating the gradation conversion curve shown in FIG. 6 using . In the case of such a configuration, operations related to gradation conversion can be performed simply by inputting the white density (D _W ) and black density (D _B ) for each input original image and the matching selection symbol for the gradation correction curve. becomes significantly simpler. Effects of the Invention By using the present invention as described above, it is possible to draw an extremely rough gradation correction curve instruction on the normalized output density versus normalized input density coordinate axis plane, which conventionally relied on the operator's intuition and experience. By specifying the coordinate values of multiple points on an arbitrary tone correction curve, it is possible to specify the gray scale conversion of an image that matches human visual characteristics, but also by specifying the coordinate values indicating various tone correction curves. By simply inputting , it is now possible to easily perform extremely wide-ranging gradation conversion processing, including density extraction processing, stepped density division processing, etc. of the input image. For example, it is particularly effective when creating artificial images at high speed, as seen in recent computer graphics.

[Brief explanation of drawings]

Fig. 1 is a block diagram of a gradation converter in a conventional image signal processing device, Fig. 2 is an input/output image signal characteristic diagram of a range amplifier, and Fig. 3 is an input/output image signal characteristic diagram of a conventional gradation converter. , FIG. 4 is a block diagram of the gradation conversion section in the apparatus used in the image signal processing method of the present invention, FIG. 5 is a diagram showing an example of the gradation correction curve to be input from the console, and FIG. Figure 7 shows a gradation conversion curve obtained from the gradation conversion curve in Figure 5.
Figure 8 is a flowchart showing the first processing method for converting the gradation correction curve in Figure 5 to the gradation conversion curve in Figure 6;
Figures a to c are diagrams showing examples of storage area arrangement in the flowchart of Figure 7, Figure 9 is a timing diagram for transferring gradation conversion data based on Figure 4, and Figure 10 is an image signal of the present invention. 11 is a flowchart showing a processing method for converting a gradation correction curve into a gradation conversion curve in the processing method. FIG. 11 is a diagram showing an example of the arrangement of storage areas in the embodiment of FIG. 10. 1...Preamplifier, 2...Range amplifier, 3...
...Grade converter, 4...Operation console, 11...A/D converter, 12...Address selector, 13...Grade conversion memory, 14...Microprocessor, 15
...Operation console, 16, 21...Interface,
17, 18, 20...Storage area, 22...Address counter, 23...Timing generator, _v1 ...Input image signal, _c1 ...White density (D _W ), _c2 ...Black density (D _B ) _c3 to _c5 ... Control signal for gradation correction curve, _v5 ...
...Digital image signal, v ₈ ...Address data line, c ₉
...Gradation conversion data line.

Claims (1)

[Scope of Claims] 1. Amplifying means for amplifying the amplitude of an analog read image signal proportional to the amount of reflected light of a document image within a predetermined amplitude range, and an analog/digital device for converting the output of the amplifying means into a digital input image signal. a conversion means;
an input means for inputting the coordinate values of a plurality of points on the coordinate axis plane of the normalized input density versus the normalized output density, as well as white density and black density; gradation correction curve creation means for creating a gradation correction curve; and gradation for creating a gradation conversion curve in a coordinate axis plane of output image signal level versus input image signal level from the gradation correction curve created by the gradation correction curve creation means. An image signal comprising: a tone conversion curve creating means; and a tone converting means for converting the tone of a digital input image signal, which is an output of the analog/digital converting means, based on the tone conversion curve of the tone conversion curve creating means. For the processing device, the procedure for creating gradation conversion curve data in the gradation conversion curve creation means is such that the digital input image signal is N bits, the maximum digital output image signal level is M, and the data is input from the input means. When the values of white density and black density are respectively D _W and D _B , and the coordinate values of multiple points P _o on the gradation correction curve are (x _o , y _o ) (N, n are natural numbers), Digital input image signal v _iI
The coordinate axes (x _o ′, y _o ′) of point P _o ′ on the digital output image signal v _pI coordinate plane are x ₁ ′=2 ^N −1/1−10 ^DW-DB {1−10( ^DW-DB )×xI} y ₁ ′=M× _{y 1} (where = 1 to ⁿ ) point
Compared with the coordinate value x _o ′ of P _o ′, x _J ′v _iI x _J+1 ′ (J=
0 to n), and from F _J indicating the straight line between point P _J ′ and point P _J+1 ′, use the relational expression v _pI = F _J (v _iI ) to obtain the digital input image signal v _iI An image signal processing method that obtains a gradation-modulated digital output image signal v _pI for.